FINAL PROJECT: Abstract and Reader's Reponse > Sampling Techniques in Data Analysis
L -- this looks like a good plan. Do you have sections titled now? You might want to refer your readers to additional sources, at the end of each section. Recall, that you do not need to teach the reader EVERY thing about sampling. You provide a guide, with referrals.
For your difficult reader, consider that you must advise them to pick a method, and then be aware of the bias or limitation. This might be a good way to conclude the document.
For your difficult reader, consider that you must advise them to pick a method, and then be aware of the bias or limitation. This might be a good way to conclude the document.
May 7, 2015 |
Marybeth Shea
Marybeth Shea
Sampling methods are important and useful in data analysis, but choosing a right sampling methods for a given situation may be a daunting task especially for those who do not have much statistics background. When possible, investigating every element in the population is the best way. However, in real life, the size of the investigated population may be large but the time and resources we possess are limited, sampling methods can be powerful since we can investigate certain selected group from population to understand the population as a whole. When sampling methods are correctly selected and applied, we should have a fair understanding of the whole population since the standard error is small. Sampling methods can be divided into two main categories: probability sampling and non-probability sampling. The characteristic of probability sampling is that the every element in the population has a probability so that they can be randomly selected. Non-probability sampling, however, does not use random selection. Elements in the sample are selected according to the people who are conducting the survey and therefore the selection may be biased. Probability sampling methods are more often used since their standard errors can be computed so that people can know how accurate the estimator is. The standard error in non-probability sampling usually cannot be computed since not all elements in investigated population are accessible, we do not even know if such error exits.
WC = 232
READER'S PROFILE:
I imagine a reader who does not have any statistics background but needs to conduct a survey.
READER'S RESPONSE:
I am recently conducting a survey and need to learn more about sampling methods. The sampling methods you discussed may be good representatives among them since I do not have enough knowledge to judge on you. In addition, some concepts in the descriptions of those sampling methods may be confusing since they need some prerequisites of statistics. In particular, calculating sample size and standard errors involves some terms that may be easy for people who already know statistics but may not be evident for people like me.